Comparison of the Forest Aboveground Biomass Inversion Methods Based on LiDAR
Stepwise regression (SR),support vector regression (SVR) and random forest (RF) were used to establish the relationship between biomass and LiDAR data.Forty variables including 20 percentile height variables and 20 point cloud density variables were used in the inversion process as independent variables.An evaluation of the regression results of these three methods was analyzed.A total of 80 field measured forest plots were used as reference dataset,including 41 needle-leaved forest (NLF) plots,19 broad-leaved forest (BLF) plots and 21 needle-broad-leaved mixed forest (NBMF) plots.An evaluation of the regression results of these three methods was analysed.The results showed that: 1) The performance of RF is better than that of SR and SVR obviously.The performance of support vector regression is better than that of stepwise regression and 2) for different forest types,NLF has best estimation accuracy,followed by BLF and NBMF.
Yunfei Wang Yong Pang Kaiguang Zhao Zengyuan Li Kairui Zhao
Institute of Forest Resource Information Techniques, Chinese Academy of Forestry;College of Forestry Institute of Forest Resource Information Techniques, Chinese Academy of Forestry Ohio Agricultural and Research Development Center, School of Environment and Natural Resources, The
国际会议
北京
英文
55-62
2013-10-09(万方平台首次上网日期,不代表论文的发表时间)